Intensive Care MedDOI 10.1007/s00134-006-0188-4 SPECIAL ARTICLE

Nicole BoluytCasper W. BollenAlbert P. BosJoke H. KokMartin Offringa

Fluid resuscitation in neonatal and pediatrichypovolemic shock: a Dutch Pediatric Societyevidence-based clinical practice guideline

Received: 16 August 2005Accepted: 12 April 2006

© Springer-Verlag 2006

This article is discussed in the editorialavailable at: http://dx.doi.org/10.1007/s00134-006-0189-3

The work was funded by the ‘PracticeGuidelines Program’ at the AcademicMedical Centre in Amsterdam.

N. Boluyt (�) · M. OffringaAcademic Medical Centre, Department ofPediatric Clinical Epidemiology, EmmaChildren’s Hospital,22700, 1100 DD Amsterdam,The Netherlandse-mail: n.boluyt@amc.uva.nlTel.: +31-20-5668822Fax: +31-20-6912683

C. W. BollenUniversity Medical Centre Utrecht,Department of Pediatric Intensive Care,Wilhelmina Children’s Hospital,85090, 3508 AB Utrecht, The Netherlands

J. H. Kok · M. OffringaEmma Children’s Hospital, AcademicMedical Centre, Department ofNeonatology,22700, 1100 DD Amsterdam,The Netherlands

A. P. BosEmma Children’s Hospital, AcademicMedical Centre, Department of PediatricIntensive Care,22700, 1100 DD Amsterdam,The Netherlands

Abstract Objective: To developa clinical practice guideline that pro-vides recommendations for the fluid,i.e. colloid or crystalloid, used forresuscitation in critically ill neonatesand children up to the age of 18 yearswith hypovolemia. Methods: Theguideline was developed througha comprehensive search and analysisof the pediatric literature. Recommen-dations were formulated by a nationalmultidisciplinary committee involv-ing all stakeholders in neonatal andpediatric intensive care and werebased on research evidence from the

literature and, in areas where theevidence was insufficient or lacking,on consensus after discussions inthe committee. Results: Because ofthe lack of evidence in neonates andchildren, trials conducted in adultswere considered. We found severalrecent meta-analyses that show excessmortality in albumin-treated groups,compared with crystalloid-treatedgroups, and one recent large ran-domized controlled trial that foundevidence of no mortality difference.We found no evidence that syntheticcolloids are superior to crystalloidsolutions. Conclusions: Given thestate of the evidence and taking allother considerations into account, theguideline-developing group and themultidisciplinary committee recom-mend that in neonates and childrenwith hypovolemia the first-choicefluid for resuscitation should beisotonic saline.

Keywords Practice guideline ·Systematic review · Shock · Fluidtherapy · Crystalloid · Colloid ·Albumin · Child

Introduction

Hypovolemia is the most common cause of circulatoryfailure in children. When inadequate tissue perfusion isnot recognized and treated during a narrow window of op-portunity, critical tissue hypoxia may develop, leading toa cascade of events resulting in multiple organ failure and

death. For this reason the concept of early goal-directedtherapy was introduced by Rivers [1]. The authors showthat this concept provides significant benefits with respectto outcome in adult patients.

The first step in the treatment of hypovolemic shockis adequate fluid resuscitation with either a crystalloid ora colloid solution. Pediatric advanced life-support guide-

lines recommend up to 60 ml/kg fluid resuscitation dur-ing treatment of hypovolemic and septic shock [2]. Fordecades there has been controversy over the relative ben-efits of crystalloid versus colloid solutions for fluid resus-citation of hypovolemic patients [3]. Since the publicationof two systematic reviews in 1998 in the British MedicalJournal [4, 5] and the Cochrane Library [6, 7], the debatehas intensified. These reviews of clinical trials conductedpredominantly in adults consistently show an excess mor-tality of around 6% in critically ill patients who receivedhuman albumin solutions in comparison with patients whoreceived a crystalloid.

In the Netherlands, the current uncertainty has resultedin the plea for a national evidence-based guideline in thepediatric age group. The goal of this guideline is to definethe current optimal choice of fluid used for treatment ofneonates and children with circulatory failure due to hypo-volemia, and to attain more uniformity in clinical practice.Recently the final version of this guideline was issued; itis now recommended and endorsed by the Dutch PediatricSociety.

In this paper we describe the methods and results ofthis guideline developmental process, including (1) a sur-vey questionnaire to assess pre-guideline volume replace-ment strategies on neonatal and pediatric intensive careunits in the Netherlands; (2) a systematic review of all ran-domized controlled trials on fluid resuscitation in hypov-olemic neonates and children; (3) all systematic reviewson fluid resuscitation in hypovolemic adults; (4) other con-siderations taken into account to reach consensus in ournational committee; and (5) the final guideline recommen-dations.

Methods

Figure 1 outlines the guideline development process.To develop this clinical practice guideline, we formed

a guideline development group consisting of five members.A national multidisciplinary committee was formed com-prising 29 members of all relevant disciplines and stake-holders (see Appendix).

As one of the goals of this guideline was to achievemore uniformity in clinical practice policy, we developeda questionnaire to first investigate the current Dutch pedi-atric practice of fluid solutions used for volume resuscita-tion. The questionnaire was sent to all directors of neona-tal (n = 10) and pediatric intensive care units (n = 8) in theNetherlands.

Based on the results of this questionnaire, the followingquestions were formulated: (1) What type of fluid solutionshould be used for initial resuscitation of hypovolemia inneonates and children? (2) What is the optimal amount offluid to be given and at what infusion rate? (3) What arethe possible side effects related to each type of fluid, suchas hypernatremia and peripheral edema?

Fig. 1 Guideline development process

Studies were identified by sensitive computerizedsearches of Medline (1966–2000), Embase (1988–2000)and the Cochrane Library, with the help of a clinicallibrarian. In addition, reference lists of all availablearticles were reviewed to identify additional citations

not found in the computerized search. Studies writtenin English, French, German and Dutch were eligible forinclusion. We searched for guidelines, systematic reviews,meta-analyses and randomized controlled trials on volumeresuscitation in critically ill neonates, children and adultswith hypovolemia due to septic shock, trauma, dehydra-tion, hemorrhage and post-cardiac surgery. With regardto question 2 we searched for studies comparing differentvolumes and rates of infusion in critically ill children.In October 2005 we repeated our literature search andlooked for additional systematic reviews and randomizedcontrolled trials in neonates and children. Studies wereincluded only when they concerned clinically relevantoutcomes, i.e. mortality and major morbidity such aspulmonary edema, length of stay in hospital, intraventric-ular hemorrhage or impaired neurodevelopment. Furtherdetails on the search strategy can be requested from theauthors.

Each study was assessed independently for itsmethodological quality by two investigators usingcritical appraisal forms originally published in JAMA(http://ugi.usersguides.org/usersguides/hg/hh_start.asp).Disagreement between these raters was resolved byconsensus. Each article was assigned a ‘level of evidence’(Table 1), which in turn influenced the ‘grade of recom-mendation’ (Table 2). These grades of recommendationoriginate from the US Agency for Health Care Policy andResearch [8].

The committee met on two separate occasions to for-mulate the final recommendations. In the event that therewas not enough evidence or the evidence was of poor qual-ity, the recommendations were based on consensus after

1a Evidence obtained from meta-analysis of randomized controlled trials1b Evidence obtained from at least one randomized controlled trial2a Evidence obtained from at least one well-designed controlled study without randomization2b Evidence obtained from at least one other type of well-designed quasi-experimental study*3 Evidence obtained from well-designed non-experimental descriptive studies, such as

comparative studies, correlation studies and case studies4 Evidence obtained from expert committee reports or opinions and/or clinical experiences

of respected authorities

* Refers to a situation in which implementation of an intervention is outside of the control of the inves-tigators, but an opportunity exists to evaluate its effect

Table 1 Classification ofevidence levels

A Requires at least one randomized controlled trial as part of a body of literature of overallgood quality and consistency addressing specific recommendation(Evidence levels 1a, 1b)

B Requires the availability of well-conducted clinical studies but no randomized clinical trialson the topic of recommendation(Evidence levels 2a, 2b, 3)

C Requires evidence obtained from expert committee reports or opinions and/or clinicalexperiences of respected authorities. Indicates an absence of directly applicable clinicalstudies of good quality.(Evidence level 4)

Note: These classifications of types of evidence and the corresponding grades of recommendation orig-inate from the US Agency for Health Care Policy and Research [8]

Table 2 Classification of gradesof recommendations

discussion in the committee and finally by show of hands.The following ‘other considerations’ for reaching consen-sus were taken into account: (a) potential side effects ofcolloids and crystalloids, (b) current insight in pathophys-iological mechanisms and their impact on the applicabilityof evidence from adults to children and neonates and (c)costs. All three aspects will be discussed below.

After a draft guideline was released it was pilotedamong end-users in the national multidisciplinary commit-tee’s hospital wards; feedback was received and includedin the final version of the guideline. A comprehensivetechnical report can be obtained from the authors.

Results

Questionnaire

The response to the questionnaire was 10/10 (100%) ofthe neonatologists and 7/8 (88%) of the pediatric inten-sivists. First-choice fluid for volume resuscitation was in50% a crystalloid and in 50% a colloid solution for bothneonatologists and pediatric intensivists. The neonatolo-gists used human albumin as a priority colloid, and the pe-diatric intensivists predominantly used a synthetic colloid,e.g. Gelofusine.

Literature

We did not identify any evidence-based guidelines on thistopic. Our search for systematic reviews, meta-analyses

and randomized controlled trials identified 93 citations,of which 65 met the inclusion criteria. Twelve articlesconcerned children or neonates. Included systematicreviews and randomized trials are summarized in Table 3and Table 4, respectively.

Question 1: What type of fluid solution should be usedfor initial resuscitation of hypovolemia?

Premature and full term neonates We identified onemeta-analysis by Kirpalani, which was excluded because itdoes not analyze hypovolemic neonates separately [9]. Weidentified five randomized controlled trials [10–14]. Fourstudies did not meet our inclusion criteria: three focused ongiving fluids prophylactically after birth [10–12] and oneon giving albumin when there was hypoalbuminemia [13].Only the study by So [14] met our inclusion criteria. Theinvestigators undertook a randomized controlled trial tostudy the efficacy of a colloid (i.e. 5% albumin) versusa crystalloid (i.e. isotonic saline) in the treatment of 63 hy-potensive preterm neonates. Outcomes, as assessed by thenumber of infants that died (RR 1.36; 95% CI 0.69–2.66),chronic lung disease (RR 0.48; 95% CI 0.13–1.87) or intra-ventricular hemorrhage (RR 1.52; 95% CI 0.91–4.87), didnot differ significantly between the groups, but the wideconfidence intervals indicate that the study was underpow-ered.

Older children Our search identified six random-ized trials [15–20]. Four studies were excluded: oneof them assessed albumin supplementation for hypoal-buminemia [16], another assessed hypertonic salinein neurotrauma [18], the third included children withsevere malaria and metabolic acidosis, not hypovolemiaper se [19], and the study by Ngo [17] used surrogateendpoints. Together with the recently published Willstrial [20], the Ngo study is the only large randomizedcontrolled trial in children. We will briefly discuss theresults of this trial, although it does not meet our inclusioncriterion on clinical relevant outcomes.

Ngo performed a randomized, double-blind trial com-paring the efficacy of four different fluid regimens (dextran70, 3% gelatin, lactated Ringer’s, and isotonic saline) inthe initial management of dengue shock syndrome (DSS)in 222 children aged 1–15 years. The primary outcomemeasures were the initial pulse pressure recovery timeand the occurrence of subsequent episodes of shock.Secondary outcome measures included the developmentof ‘any complication’ of fluid therapy. There were nodeaths in any of the groups and there were no differencesin the ‘reshock’ rate among the four groups. Six childrenhad allergic reactions after colloid therapy (five receivedgelatin and one dextran), defined as fever and chills. Onechild in the gelatin group had severe epistaxis and anotherchild in the dextran group a large hematoma at a site ofminor trauma. In this study no clear benefits of any oneof the four fluids in improving these surrogate endpointscould be demonstrated. Normal saline performed as wellas the colloid solutions. Ta

ble

3S

umm

ary

ofin

clud

edsy

stem

atic

revi

ews

com

pari

ngth

eef

fect

ofan

yfl

uid

solu

tion

wit

hot

her

flui

dso

luti

ons

Stud

yY

ear

Age

Inte

r-C

ompa

riso

nO

utco

mes

Ade

quat

eQ

ualit

yD

escr

iptio

nFo

cuse

dD

escr

iptio

nD

escr

iptio

nV

alid

Ass

essm

ent

RR

Lev

elgr

oup

vent

ion

sear

chas

sess

men

tof

artic

les

clin

ical

ofse

lect

ion

ofda

tast

atis

tical

of(9

5%C

I)a

ofst

rate

gyof

RC

Ts

ques

tion

proc

edur

eex

trac

tion

pool

ing

hete

roge

neity

evid

ence

ofR

CT

s

Vel

anov

ich

[24]

1989

Adu

ltC

ollo

idC

ryst

allo

idD

eath

No

No

Yes

Yes

No

No

Yes

No

NG

NA

Bis

onni

[21]

1991

Adu

ltC

ollo

idC

ryst

allo

idD

eath

No

No

No

Yes

No

No

No

No

NG

NA

Schi

erho

ut[5

]19

98A

dult

Col

loid

Cry

stal

loid

Dea

thY

esY

esY

esY

esY

esY

esY

esY

es1.

521a

(1.0

8-2.

13)

Cho

i[23

]19

99A

dult

Col

loid

Cry

stal

loid

Dea

th,

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

0.86

1apu

lmon

ary

(0.6

3-1.

17)

edem

aB

unn

[22]

2000

Adu

ltC

ollo

idC

ollo

idD

eath

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

0.99

1a(0

.69-

1.42

)A

lder

son

[7]

2002

Adu

ltA

lbum

inN

oal

bum

inD

eath

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

1.46

1a(0

.97-

2.22

)

NG

,not

give

n;N

A,n

otap

plic

able

aM

orta

lity

RR

inhy

povo

lem

icpa

tient

s

Tabl

e4

Sum

mar

yof

incl

uded

pedi

atri

cra

ndom

ized

tria

lsco

mpa

ring

the

effe

ctof

any

flui

dso

luti

onw

ith

othe

rfl

uid

solu

tion

s,an

dth

ere

cent

SA

FE

tria

l

Stud

yY

ear

Age

No

ofIn

ter-

Com

pari

son

Out

com

eA

dequ

ate

Suffi

cien

tB

lindi

ngB

lindi

ngB

lindi

ngof

Patie

nts

Equ

alR

R(9

5%C

I)L

evel

ofgr

oup

patie

nts

vent

ion

rand

om-

follo

w-u

pof

ofef

fect

prog

nost

icco

ncom

itant

mor

talit

yev

iden

ceiz

atio

npa

tient

str

eatm

ent

asse

ssm

ent

equa

ltr

eatm

ent

Bol

dt[1

5]19

93C

hild

30A

lbum

inH

ES

Dea

th,

Yes

No

No

No

No

No

No

No

deat

hsan

d1b

pulm

onar

yno

pulm

onar

yed

ema

edem

aoc

curr

edSo

[14]

1997

Chi

ld63

Alb

umin

Cry

stal

loid

Dea

th,

Yes

Yes

No

No

Yes

Yes

Yes

1.36

1bC

LD

∗ ,(0

.69–

2.66

)IV

H∗∗

SAFE

Stud

y20

04A

dult

6997

Alb

umin

Isot

onic

Dea

thY

esY

esY

esY

esY

esY

esY

es0.

991b

Inve

sti-

salin

e(0

.91–

1.09

)ga

tors

[26]

Will

s[2

0]20

05C

hild

512

Rin

ger’

sla

ctat

eN

o.of

Yes

Yes

Yes

Yes

Yes

Yes

Not

stat

edN

otgi

ven

1bor

6%de

xtra

nof

days

70(a

collo

id)

orin

hosp

ital

6%hy

drox

yeth

ylst

arch

(aco

lloid

)

∗C

LD

=ch

roni

clu

ngdi

seas

e∗∗

IVH

=in

trav

entr

icul

arhe

mor

rhag

e

Boldt conducted a randomized study in 30 childrenless than 3 years of age undergoing cardiac surgery [15].They were randomly assigned to receive albumin orlow-molecular-weight hydroxyethyl starch solution (6%LMW-HES). No deaths occurred and none of the childrenhad signs of pulmonary edema, but the wide confidenceintervals indicate that the study was underpowered.

Recently, Wills conducted a large randomized con-trolled trial in 512 children with dengue shock syndrome,comparing three resuscitation fluids; Ringer’s lactate, 6%dextran 70 and 6% hydroxyethyl starch [20]. A total of383 children with moderately severe shock were randomlyassigned to receive one of the three fluids and 129 childrenwith severe shock to receive one of the colloids. Theprimary outcome was the requirement for supplementalintervention with rescue colloid. Secondary outcomemeasures were ‘time taken to achieve initial and sustainedcardiovascular stability’ and ‘number of days in hospital’.The relative risk of a requirement for rescue colloid was1.08 (95% CI 0.78–1.47) among children with moderateshock who received Ringer’s lactate as opposed to eitherof the colloid solutions. There were no differences in thetime to final cardiovascular stability or the number of daysin the hospital.

Adults Because of the lack of randomized controlledtrials in children, we included trials conducted in criticallyill adults. We found seven meta-analyses [4–7, 21–25] Thesystematic review by Wilkes was excluded because it didnot analyze hypovolemic patients separately [25]. Recentlya large randomized controlled trial, the Saline versus Albu-min Fluid Evaluation (SAFE) Study, was published [26].Because of poor methodological quality we will not dis-cuss the two older systematic reviews that compare col-loids with crystalloids in fluid resuscitation here [21, 24].Instead, the results of the remaining four reviews and theSAFE Study will be discussed below.

All four remaining systematic reviews are of goodmethodological quality (Table 3). The reviews by Aldersonand Schierhout comparing colloids with crystalloids bothshow a 6% increase in mortality in the group receivingalbumin [5, 6].

The meta-analysis by Choi [23] observed no differencebetween crystalloid and colloid resuscitation with respectto pulmonary edema [pooled RR 0.84 (0.25–2.45)] andmortality [pooled RR 0.86 (0.63–1.17)]. However, thepower of the aggregated data was insufficient to detectsmall but potentially clinically important differences.Subgroup analysis showed a statistically significant re-duction in mortality rate in trauma in favor of crystalloidresuscitation (RR 0.39, 95% CI 0.17–0.89).

The meta-analysis by Bunn did not show evidence thatone colloid solution is more effective or safe than anyother: albumin vs. HES: RR 1.17 (0.91–1.50), albuminvs gelatin: RR 0.99 (0.69–1.42) [22]. Again, in this studythe confidence intervals are wide and the results do notexclude clinically relevant differences among colloids.

The SAFE Study investigators conducted a large ran-domized controlled trial in adult patients who had beenadmitted to the ICU and required fluid administration tomaintain or increase intravascular volume [26]. A total of6,997 patients were randomly assigned to receive 4% al-bumin or isotonic saline. The relative mortality risk amongpatients assigned to receive albumin compared with thoseassigned saline was 0.99 (95% CI 0.91–1.09). The authorsconclude that there is evidence of no difference in mortal-ity rate.

It is possible that there are certain subgroups whereeither colloids or crystalloids are more effective. Thesefour meta-analyses comprised different patient categories,namely trauma, hypoalbuminemia, hypovolemia, sepsis,burns, and cardiopulmonary surgery. There was no evi-dence of advantage of colloids over crystalloids in anyof these different indication subgroups, although in mostcases the results were inconclusive because of lack ofstatistical power. Especially in patients after trauma orburns, albumin appeared to be associated with increasedmortality and increased ventilator requirement. In traumapatients, the relative mortality rates when comparingalbumin and saline were between 0.98 and 5.88. TheSAFE Study included post-hoc subgroup analyses. Therelative risk of death in the albumin group compared withpatients in the saline group was 1.36 (95% CI 0.99–1.87)for trauma patients, 0.87 (95% CI 0.74–1.02) for patientswith sepsis and 0.93 (95% CI 0.61–1.41) for patients withrespiratory distress syndrome. It must, however, be notedthat in large studies such subgroup differences frequentlyoccur by chance, and common wisdom holds that hy-potheses generated in this manner should be evaluated inspecifically designed and appropriately powered futurestudies.

Question 2: What is the optimal amount to give and atwhat infusion rate?

We identified only one study by Carcillo about therole of early and rapid fluid resuscitation in children withseptic shock [27]. Included were 34 children (median age13.5 months) with septic shock who all required vaso-pressor and/ or inotropic support. Therapeutic decisionswere left to the attending staff in this observational study.At 1 h and 6 h, respectively, group 1 (n = 14) received11 ± 6 and 71 ± 29 ml/kg (mean ± SD) of fluid; group 2(n = 11) received 32 ± 5 and 108 ± 54 ml/kg of fluid; andgroup 3 (n = 9) received 69 ± 19 and 117 ± 29 ml/kg offluid. Fluids used were 5% albumin, fresh frozen plasma,cryoprecipitate, isotonic saline and lactated Ringer’ssolution. Details on which patient received which fluidwere not given in the paper. Survival in group 3 (8 of9 patients) was significantly better than in group 1 (6 of14 patients) or group 2 (4 of 11 patients). The authorsconcluded that rapid fluid resuscitation in excess of40 ml/kg in the first hour following emergency departmentpresentation of children with septic shock is associatedwith improved survival. However, since the treatment

groups were assigned non-randomly and the choice oftreatment was based on clinical criteria that were deter-mined by individual physicians, and since no adjustmentswere made for co-interventions like the use of inotropics,these observations cannot prove cause and effect. Level ofevidence: 2b.

There is no evidence in children and neonates with hy-povolemia not caused by septic shock about the optimalvolume to be used and the velocity of fluid resuscitation.The efficacy of fluid replacement depends on the existingmicrovascular pressures, the compliance of the interstitialspace and the permeability of the microvascular barrier.Hence both the volume and velocity of fluid resuscitationshould be determined individually.

Question 3: What are the possible side effects relatedto the type of fluid used, such as hypernatremia and pe-ripheral edema?

Hypernatremia Hypernatremia is thought by someto play a role in the development of intraventricularhemorrhage in neonates. There is a concern that by givingisotonic saline hypernatremia will be induced, because ofa diminished renal sodium clearance potential in the firstweek of life. One of the outcomes assessed in the studyby So was the serum sodium concentration [14]. Theinvestigators included 63 hypotensive premature infants,randomly assigned to receiving a colloid or crystalloidsolution. They could not find a significant difference inmean sodium concentration, or the rate of intraventricularhemorrhage or mortality, between the two groups. Thislack of difference observed may be partly due to smallnumbers of patients. Furthermore, in most countries 5%albumin contains almost the same amount of sodium asisotonic saline (145 mmol/l vs. 154 mmol/l).

Peripheral edema Both the use of colloids and crys-talloids is accompanied by the occurrence of peripheraledema [16, 28–31]. More important to know is howquickly such edema can be mobilized and whether thereare any clinically relevant consequences. In our sensitivesearches for evidence, no studies that give answers tothese questions could be found.

Other considerations taken into account at theconsensus meeting

Pathophysiology and applicability of the evidence fromadults in neonates and children

If the alveolo-capillary membrane is intact, the lungs arewell protected against a drop of colloid osmotic pressure(or hypoalbuminemia); if the membrane is damaged, theinfusion of colloid aimed at increasing the colloid osmoticpressure is illusive, since the colloids leak into the intersti-tium and could even amplify the pulmonary and peripheraledema.

1. In neonates and children with hypovolemia the first-choice fluid for initial resuscitationis isotonic saline (Grade A)*

2. When large amounts of fluids are required (e.g. sepsis), it is possible to use a synthetic colloidbecause of its longer duration in the circulation (Grade C)

3. The initial fluid volume should be 10–20 ml/kg and repeated doses should be basedon individual clinical response (Grade C)

* In adults there is grade A evidence. For reasons described in this paper this evidence is thought to beapplicable to neonates and children.

Table 5 The guideline’srecommendations

Although most of the evidence in this field comesfrom studies in critically ill adults, the results of theseinvestigations do not differ from those studies that havebeen conducted in critically ill neonates and children;there is just more information on adults. All studies failto show an advantage of colloids over crystalloids whensurvival rates are considered. In addition, there is no clearevidence whether colloids or crystalloids confer benefitin certain subgroups of patients with shock, i.e. septic,hemorrhagic or hypovolemic. The reported increasedmortality rate due to the use of albumin could be ex-plained by the distribution of albumin across the capillarymembrane, a process that is accelerated in criticallyill patients [32]. Increased leakage of colloids into theextra vascular spaces might reduce the oncotic pressuredifference across the capillary wall, making edema morelikely. Yet, given the results of the SAFE Study, thismechanism may exist, but it may be less important inclinical practice than we thought previously. Although thedistribution of body fluids in the neonate differs from thatin adults, there is currently no reason to believe – eitherfrom pathophysiological theory or empirical evidence– that once a capillary leak exists, albumin would bebeneficial in children.

It is, however, possible that in contrast to albumin,which is a small molecule (60 kDa), synthetic colloidswith larger molecules (HES, 200 kDa) do not leak intothe interstitial space. Yet, so far there is no evidencethat some colloids are more effective or safer thanothers, when clinically relevant outcomes are consid-ered.

Isotonic saline is distributed equally throughout theextracellular space. Because the extracellular fluid isone-fourth intravascular and three-fourths interstitial,only one-fourth of the infusate remains intravascular.For this reason some members of the multidisciplinaryconsensus group believed that it might be more efficientto use a synthetic colloid after initial crystalloids inpatients with refractory hypovolemia and significanthemodynamic problems (e.g. those with severe sepsis).However, the SAFE Study showed that the ratio of thevolume needed to maintain stable circulation of albu-min to the same volume of isotonic saline administeredwas only 1.4. Again, we currently have no reason tobelieve that this ratio would be different in neonates andchildren.

Costs

Colloid solutions are much more expensive than crys-talloid solutions: 1 l of albumin currently costs around140 Euro (152 US$), 1 l of HES costs 25 Euro (27 US$)and 1 l of isotonic saline costs 1.5 Euro (1.6 US$).

The committee’s recommendations

As colloids are biological products with a potential infec-tion hazard or a risk of anaphylactic reaction and becausethey are much more expensive than crystalloids, it was feltby the national multidisciplinary committee that their ben-efits over crystalloids should be proved before they wereused. Given the state of the evidence and taking all otherconsiderations into account, the guideline-developinggroup and the national multidisciplinary committee rec-ommend that in neonates and children with hypovolemiathe first-choice fluid for initial resuscitation should beisotonic saline. When large amounts of fluids are required(e.g. in sepsis), it is possible to use a synthetic colloidbecause of its longer duration in the circulation. The initialfluid volume should be 10–20 ml/kg, with repeated dosesbased on individual clinical response (Table 5).

Discussion

In adult patients the current volume of evidence on thesolutions of choice for fluid resuscitation of hypovolemicpatients is large, consistently showing no advantage ofcolloids over crystalloids. Despite this, several surveyshave shown that this evidence has not changed clinicalpractice: the majority of physicians still use colloid prod-ucts [33–35]. More interestingly, most physicians couldnot state reasons for choosing between products [34].To bridge this apparent gap between the evidence andactual clinical practice, guidelines may be needed [36,37]. Clinical practice guidelines are seen as powerfultools to achieve effective and efficient care, but havebeen demonstrated to be effective only if there is suffi-cient rigor in their method of development [38]. Havingcompleted a rigorous and objective synthesis of theevidence base, the guideline-development group mustmake what is essentially a subjective judgment on the

recommendations that can validly be made on the basisof the available evidence [39]. Subjective judgment risksthe reintroduction of bias into the process. However, inhigh-quality guideline-development processes this is notthe judgment of one individual but of a carefully composedmultidisciplinary group. An additional safeguard here isthe requirement for the guideline-development group topresent clearly the evidence on which the recommendationis based, and to make the link between the evidenceand the recommendations explicit, explaining how theevidence was interpreted [39]. Thus, the summarizedresearch evidence is only one factor in clinical guidelines,and the ‘other considerations’ are just as important inreaching consensus on the recommendations. This holdsespecially when there is insufficient research evidence.Therefore, these ‘other considerations’ should be clearlypresented in any guideline or its technical report [40].

Before the present guideline was developed there wascontroversy in the Netherlands about the fluid of choicefor volume replacement; about 50% of neonatologists andpediatric intensivists used colloids. We found that in chil-dren the volume and quality of the available research ev-idence is limited. The final recommendations are largelybased on the ‘other considerations’, in this case the poten-tial side effects of colloids and crystalloids, current insightinto pathophysiological mechanisms and their impact onthe applicability of evidence from adults to children andneonates, and costs. We involved 29 stakeholders in Dutchneonatal and pediatric intensive care practice to formulatethe recommendations and unanimously decided that iso-tonic saline should be the first choice, because it is equallyeffective, safe and up to 100 times cheaper than albumin.We realize that in other settings and countries the ‘otherconsiderations’ may play a different role and have a differ-ent bearing on decision making. As the process of devel-opment of this guideline was explicit, pediatric intensivistsand neonatologists outside the Netherlands may make theirown recommendations based on the information we haveassembled.

We believe that the introduction of this nationwideguideline for fluid resuscitation may contribute to anoptimal, cost-effective universal treatment strategy inpediatric patients with imminent circulatory failure.

Conclusion

Given the state of the evidence and taking other relevantconsiderations into account, the guideline-developinggroup and the Dutch national multidisciplinary committeerecommend that in neonates and children with hypo-volemia the first-choice fluid for resuscitation should beisotonic saline.

This guideline will be updated in July 2008

Appendix: National Multidisciplinary Committee

Neonatologists

P. Andriessen, MD, W. Baerts, MD, PhD, P.L.J. De-graeuwe, MD, W.P.F. Fetter, MD, PhD, W.B. Geven, MD,PhD, A.H.L.C. van Kaam, MD, K.D. Liem, MD, PhD,D.W.E. Roofthooft, MD, M.G.A. Verboon-Maciolek, MD,F.J. Walther, MD PhD, F. Brus, MD, PhD, W.W.M. Hack,MD, PhD, J.W.F.M. Jacobs, MD

Pediatric intensivists

P. Dahlem, MD, R.J.B.J. Gemke, MD, PhD, L.G.F.M. van‘t Hek, MD, C. Buysse, MD, G.D. Vos, MD, A.J. vanVught, MD, PhD, W. de Weerd, MD, N. van der Lelij,MD, F.B. Plötz, MD, PhD, L. Veenhuizen, MD, W.B.Vreede, MD, C.M. Walhof, MD

Others

M. Hemmink, nurse, neonatology; D. Tol, nurse, intensivecare; T.W.J. Schulpen, MD, PhD, Director, Quality Office,Dutch Pediatric Society; C.R. Lincke, MD, PhD, Chair ofthe Society’s Guideline Development Committee, DutchPediatric Society

None of the members of the study group or the nationalmultidisciplinary committee had any connections with thepharmaceutical industry or had otherwise any financial in-terest in the final recommendations.

References

1. Rivers E, Nguyen B, Havstad S,Ressler J, Muzzin A, Knoblich B, Pe-terson E, Tomlanovich M (2001) Earlygoal-directed therapy in the treatmentof severe sepsis and septic shock. NEngl J Med 345:1368–1377

2. Ceneviva G, Paschall JA, Maffei F,Carcillo JA (1998) Hemodynamicsupport in fluid-refractory pediatricseptic shock. Pediatrics 102:e19

3. Offringa M (1998) Excess mortalityafter human albumin administrationin critically ill patients. Clinical andpathophysiological evidence suggestsalbumin is harmful. BMJ 317:223–224

4. Cochrane Injuries Group AlbuminReviewers (1998) Human albuminadministration in critically ill patients:systematic review of randomisedcontrolled trials. BMJ 317:235–240

5. Schierhout G, Roberts I (1998) Fluidresuscitation with colloid or crystalloidsolutions in critically ill patients: a sys-tematic review of randomised trials.BMJ 316:961–964

6. Alderson P, Schierhout G, Roberts I,Bunn F (2000) Colloids versus crystal-loids for fluid resuscitation in criticallyill patients. Cochrane Database SystRev :CD000567

7. Alderson P, Bunn F, Lefebvre C, Li WP,Li L, Roberts I, Schierhout G (2002)Human albumin solution for resuscita-tion and volume expansion in criticallyill patients. Cochrane Database SystRev:CD001208

8. US Department of Health and HumanServices. Agency for Health CarePolicy and Research (2003) Acute painmanagement: operative or medical pro-cedures and trauma. Rockville (MD):The Agency; 1993. Clinical PracticeGuideline No.1. AHCPR PublicationNo. 92–0023. p 107

9. Kirpalani H (2001) The use of al-bumin and/or colloids in newborns:A meta-analysis. Pediatr Crit Care Med2:S14–S19

10. Northern Neonatal Nursing InitiativeTrial Group (1996) Randomised trialof prophylactic early fresh-frozenplasma or gelatin or glucose in pretermbabies: outcome at 2 years. Lancet348:229–232

11. Northern Neonatal Nursing Initiative(NNNI) Trial Group (1996) A ran-domized trial comparing the effect ofprophylactic intravenous fresh frozenplasma, gelatin or glucose on earlymortality and morbidity in pretermbabies. Eur J Pediatr 155:580–588

12. Bland RD, Clarke TL, Harden LB,Meyer JL, Ries JP, Madden WA,Crast FW, Coyer WF, Bass JW (1973)Early albumin infusion to infants at riskfor respiratory distress. Arch Dis Child48:800–805

13. Greenough A, Emery E, Hird MF,Gamsu HR (1993) Randomised con-trolled trial of albumin infusion inill preterm infants. Eur J Pediatr152:157–159

14. So KW, Fok TF, Ng PC, Wong WW,Cheung KL (1997) Randomised con-trolled trial of colloid or crystalloid inhypotensive preterm infants. Arch DisChild Fetal Neonatal Ed 76:F43–F46

15. Boldt J, Knothe C, Schindler E, Ham-mermann H, Dapper F, Hempelmann G(1993) Volume replacement with hy-droxyethyl starch solution in children.Br J Anaesth 70:661–665

16. Greenhalgh DG, Housinger TA, Ka-gan RJ, Rieman M, James L, Novak S,Farmer L, Warden GD (1995) Main-tenance of serum albumin levels inpediatric burn patients: a prospective,randomized trial. J Trauma 39:67–73

17. Ngo NT, Cao XT, Kneen R, Wills B,Nguyen VM, Nguyen TQ, Chu VT,Nguyen TT, Simpson JA, Solomon T,White NJ, Farrar J (2001) Acute man-agement of dengue shock syndrome:a randomized double-blind comparisonof 4 intravenous fluid regimens in thefirst hour. Clin Infect Dis 32:204–213

18. Simma B, Burger R, Falk M, Sacher P,Fanconi S (1998) A prospective, ran-domized, and controlled study of fluidmanagement in children with severehead injury: lactated Ringer’s solutionversus hypertonic saline. Crit Care Med26:1265–1270

19. Maitland K, Pamba A, English M,Peshu N, Marsh K, Newton C, Levin M(2005) Randomized trial of volumeexpansion with albumin or saline inchildren with severe malaria: prelimi-nary evidence of albumin benefit. ClinInfect Dis 40:538–545

20. Wills BA, Nguyen MD, Ha TL,Dong TH, Tran TN, Le TT, Tran VD,Nguyen TH, Nguyen VC, Step-niewska K, White NJ, Farrar JJ (2005)Comparison of three fluid solutionsfor resuscitation in dengue shocksyndrome. N Engl J Med 353:877–889

21. Bisonni RS, Holtgrave DR, Lawler F,Marley DS (1991) Colloids versuscrystalloids in fluid resuscitation: ananalysis of randomized controlled trials.J Fam Pract 32:387–390

22. Bunn F, Alderson P, Hawkins V (2001)Colloid solutions for fluid resusci-tation. Cochrane Database Syst Rev:CD001319

23. Choi PT, Yip G, Quinonez LG, Cook DJ(1999) Crystalloids vs. colloids in fluidresuscitation: a systematic review. CritCare Med 27:200–210

24. Velanovich V (1989) Crystalloidversus colloid fluid resuscitation:a meta-analysis of mortality. Surgery105:65–71

25. Wilkes MM, Navickis RJ (2001) Patientsurvival after human albumin adminis-tration. A meta-analysis of randomized,controlled trials. Ann Intern Med135:149–164

26. Finfer S, Bellomo R, Boyce N, French J,Myburgh J, Norton R (2004) A com-parison of albumin and saline for fluidresuscitation in the intensive care unit.N Engl J Med 350:2247–2256

27. Carcillo JA, Davis AL, Zaritsky A(1991) Role of early fluid resuscita-tion in pediatric septic shock. JAMA266:1242–1245

28. Ernest D, Belzberg AS, Dodek PM(1999) Distribution of normal salineand 5% albumin infusions in septicpatients. Crit Care Med 27:46–50

29. Lucas CE, Weaver D, Higgins RF,Ledgerwood AM, Johnson SD, Bouw-man DL (1978) Effects of albuminversus non-albumin resuscitation onplasma volume and renal excretoryfunction. J Trauma 18:564–570

30. Pockaj BA, Yang JC, Lotze MT,Lange JR, Spencer WF, Steinberg SM,Topalian SL, Schwartzentruber DJ,White DE, Rosenberg SA (1994)A prospective randomized trial eval-uating colloid versus crystalloidresuscitation in the treatment of thevascular leak syndrome associated withinterleukin-2 therapy. J Immunother15:22–28

31. Prien T, Backhaus N, Pelster F,Pircher W, Bunte H, Lawin P (1990)Effect of intraoperative fluid admin-istration and colloid osmotic pressureon the formation of intestinal edemaduring gastrointestinal surgery. J ClinAnesth 2:317–323

32. Berger A (1998) Why albumin may notwork. BMJ 317:240

33. Boldt J, Lenz M, Kumle B, Papsdorf M(1998) Volume replacement strategieson intensive care units: results froma postal survey. Intensive Care Med24:147–151

34. Miletin MS, Stewart TE, Norton PG(2002) Influences on physicians’choices of intravenous colloids. Inten-sive Care Med 28:917–924

35. Soderlind M, Salvignol G, Izard P,Lonnqvist PA (2001) Use of albumin,blood transfusion and intraoperativeglucose by APA and ADARPEF mem-bers: a postal survey. Paediatr Anaesth11:685–689

36. Haines A, Jones R (1994) Imple-menting findings of research. BMJ308:1488–1492

37. Woolf SH, Grol R, Hutchinson A,Eccles M, Grimshaw J (1999) Clinicalguidelines: potential benefits, limita-tions, and harms of clinical guidelines.BMJ 20;318:527–530

38. Farquhar CM, Kofa EW, Slutsky JR(2002) Clinicians’ attitudes to clinicalpractice guidelines: a systematic review.Med J Aust 177:502–506

39. Harbour R, Miller J (2001) A newsystem for grading recommendationsin evidence based guidelines. BMJ323:334–336

40. Burgers JS, van Everdingen JJ (2004)Beyond the evidence in clinicalguidelines. Lancet 364:392–393